This paper (video abstract below!) came on my radar recently, that antibiotic stewardship works in the NICU, that initiation of antibiotics is reduced, length of tx is reduced etc.

A recent UK paper is a nice example of this, although most preterm infants in this paper still recieved antibiotics

I think we are pretty restrictive with antibiotics (as illustrated by this paper), but I feel we could have a more systematic approach...

How do you tailor antibiotics, and what tools do you use?

• what makes you start?
• what makes you stop?

 

We have changed our practice in the last few years. 

General rule is if SPTL, PROM, resp support required >4hrs (or Work of breathing / Fi02 req increasing earlier) or other risk factors present at delivery, we tend to initiate IV Abx on admission (Benzyl Penicillin and Gentamicin).  If repeat CRP is normal, no growth in culture and baby is well, we tend to stop Abx prior to 24hr dose (i.e. receive 2 doses BenPen and 1 dose Gent).

Also, if prolonged raised lactates but normal neurology, we tend to septic screen and initiate IV Abx.  Same scenario, stop at 24hrs if all okay, however, we have found a number of congenital infections in these infants.  Based on ongoing clinical picture, we would consider stopping Abx at 5-10 days but if meningitis, 2-3 weeks.

For LOS, we still do jump into early septic screen and IV Abx rather quickly due to rapidness of deterioration for our extreme prem infants.  We generally use Vanc and Gent (if central lines insitu) or Fluclox / Gent (if not lines) and usually stop prior to 48hr dose if no signs sepsis.  

I spoke to an american ANP at our 99nicu-Meetup who said that introducing the Kaiser score ( https://neonatalsepsiscalculator.kaiserpermanente.org/ )had really limited the antibiotic use in their unit. If I remember it correctly they had automated it into their digital journal system (Epic?).

I'm not sure why we don't use it, is it implemented anywhere in Europe? But as CRP seems to be insufficient, even though we use it in combination with IL-6/Procalcitonin, it might be a good complement. Are there any disadvantages with Kaiser?

The talk by Rene Kornelisse at the 99nicu meetup gave some food for thought on the subject of antibiotic use, let's see if we can publish it on the 99nicu Youtube-channel soon 🙂.

We participated in an initiative The American University started for antibiotic stewardship in neonatology, and I consider it a success. It made us stop some bad habits, reconsider the duration of anti-biotherapy, and encouraged us to stop ATB instead of waiting unnecessarily for cultures in some evident cases.

Sort of on this topic but slightly different.....

I think this might be the first RCT (cluster RCT including both non-inferiority and superiority outcomes) on the EOS calculator versus 'categorical guidelines? Performed in 10 centres in the Netherlands.

https://www.thelancet.com/action/showPdf?pii=S2589-5370%2825%2900351-7

@Gustaf Lernfelt as @Stefan Johansson has highlighted in the antibiotic exposure paper (Giannoni et al 2022) Sweden has one of the lowest antibiotic exposure rates. I guess it would be a case of determining if an EOS risk calculator would reduce antibiotic exposure in Sweden, or are you already using these judiciously?

Thoughts?!

Yes, we do have a low use of antibiotics in Sweden. We did some informal looks into the Kaiser Permanent Sepsis Calculator with hypothetical cases, whether our clinical strategy would differ from the recommendation from the Calculator. It seemed that our current approach was still more restrictive compared to the Calculator, so, we have some secret sauce when it comes to anitbiotics use.

There are two recent publications about antibiotics use in Sweden, giving some raw numbers on the proportion of infants given antibiotics (roughly 2% of all NICU admissions):

• https://fn.bmj.com/content/early/2025/07/18/archdischild-2025-328944

• https://pubmed.ncbi.nlm.nih.gov/38517437/

But even with this small rate, there are substantial variations between hospitals, without different sepsis risks, so it seems possible to improve the use even further.

I am curious to hear more about this possibly even more restrictive approach! It would be interesting to know what your practice looks like!

We should invite people like Johan Gyllensvärd and Eric Giannoni to a webinar discsussion around this!

We discussed this the other day, and a colleague who recently worked in the Netherlands gave us some explanation on how they practiced at the hospital where she worked there versus here.

First of all, we all kind of use our own algorithm when it comes to risk for infection, even if it’s not put into a score. How long was it since the water broke, maternal fever, known GBS-status or was it an elective C-section?

Second to that, the region I work in use IL-6 in combination with CRP in newborns, high levels of IL-6 is a strong marker for bacterial infection. On the other hand a low IL-6 at the start of symptoms makes an infection less likely. It reacts very fast at the start of an infection, but fades down within 24 h. But at that time a CRP would be significantly high.

So let’s say that you have an infant 3 hours of age, no risk factors, with a high respiratory rate (70). IL-6 and CRP is negative, —> no AB, follow RR. If you would do a new CRP at 24 h and it would be 40 mg/L, this could still be considered possibly normal due to vaginal birth, and would prompt further follow up, but no Ab. If there were any risk factor, and the child showed symptoms we might have started antibiotics, but stopped if the CRP went down again and there were no positive blood culture in 48-72 h.

In the Netherlands where she worked everyone with a CRP above 20 mg/L would be treated with Ab. Possibly due to other organisational issues.

I know how studies have showed that CRP is CRaP, but I guess it kind of depends on how you use it. This way I find it useful, as a sign of inflammation and not patognomic for infection (depending on the level, we would say around 60 mg/L is a possible limit).